1. Domino Tilings of the Chessboard An Introduction to Sampling and Counting Dana Randall Schools of Computer Science and Mathematics Georgia Tech

2. Building short walls How many ways are there to build a 2 x n wall with 1 x 2 bricks? 2 n 2 1

3. Building short walls 2 1How many ways are there to build a 2 x n wall with 1 x 2 bricks? 2 n

4. n=1 : n=2 : n=3 : n=0 : n=4 : Building short walls

5. n=0 : n=1 : n=2 : n=3 : n=4 : The number of walls equal: f n = 1, 1, 2, 3, 5

6. Building short walls n=0 : n=1 : n=2 : n=3 : n=4 : The number of walls equal: f n = 1, 1, 2, 3, 5, 8, 13, 21,... ?

7. Building short walls n=0 : n=1 : n=2 : n=3 : n=4 : The number of walls equals: f n = 1, 1, 2, 3, 5, 8, 13, 21,... ? n 2 2 fn =fn = {

8. Building short walls n=0 : n=1 : n=2 : n=3 : n=4 : The number of walls equals: f n = 1, 1, 2, 3, 5, 8, 13, 21,... ? n 2 2 fn =fn = {

9. Building short walls n=0 : n=1 : n=2 : n=3 : n=4 : The number of walls equals: f n = 1, 1, 2, 3, 5, 8, 13, 21,... ? n 2 2 f n-1 f n-2 fn =fn = {

10. The Fibonacci Numbers The number of walls equals: f n = 1, 1, 2, 3, 5, 8, 13, 21,... f n-1 f n-2 fn =fn = { f n = f n-1 + f n-2, f 0 = f 1 = 1 f n = (φ n + (1-φ) n ) /√5, where: (“golden ratio”) φ = 1+√5 2

11. Domino Tilings Given a region R on the infinite chessboard, cover with non-overlapping 2 x 1 dominos. Where is a tiling? Do any even exist? How many tilings are there? What does a typical tiling look like? When do we stop our algorithms? Why do we care?

12. Where is a tiling? Do any exist? n n ★ Only if n is even!

13. Where is a tiling? Do any exist? The Area of R must be even n n

14. Where is a tiling? Do any exist? n n ★ There must be an equal number of black and white squares. The Area of R must be even

15. Where is a tiling? Do any exist? Is this enough? The Area of R must be even With an equal number of white and black squares

16. Where is a tiling? Do any exist? Is this enough? ?... There is an efficient algorithm to decide if R is tileable and to find one if it is. [Thurston] The Area of R must be even With an equal number of white and black squares

17. Domino Tilings Where is a tiling? Do any even exist? How many tilings are there? What does a typical tiling look like? When do we stop our algorithms? Why do we care?

18. How many tilings are there?  φ n Short 2 x n walls

19. How many tilings are there?  φ n Short 2 x n walls Aztec Diamonds n # = 1 2 8 64 n=0 n=1 n=2 n=3  n(n+1)/2 [Elkies, Kuperberg, Larson, Propp]

20. How many tilings are there?  φ n Short 2 x n walls Aztec Diamonds n  n(n+1)/2

21. How many tilings are there?  φ n Short 2 x n walls Aztec Diamonds Square n x n walls n ?  n(n+1)/2

22. How many tilings are there? Square n x n walls 2 (Area/4) < # < # < 4 (Area) <

23. How many tilings are there? #  φ n Short 2 x n walls Aztec Diamonds Square n x n walls 2 Area/4 < # < 4 Area #  n(n+1)/2

24. How many: An Algorithm Mark alternating vertical edges;

25. How many: An Algorithm Mark alternating vertical edges; Use marked tiles; Markings must line up!

26. How many: An Algorithm t1t1 t2t2 t3t3 s1s1 s2s2 s3s3

27. t1t1 t2t2 t3t3 s1s1 s2s2 s3s3 s1s1 s2s2 s3s3 t1t1 t2t2 t3t3 We want to count non-intersecting sets of paths from s i to t i. [R.]

28. How many: An Algorithm We want to count non-intersecting sets of paths from s i to t i. Let a ij be the number of paths from s i to t j. s1s1 s2s2 s3s3 t1t1 t2t2 t3t3

29. How many: An Algorithm s1s1 s2s2 s3s3 t1t1 t2t2 t3t3 Let a ij be the number of paths from s i to t j. We want to count non-intersecting sets of paths from s i to t i. 1 4 16 1 3 12 52 1 5 24 1 7 40 1 9 1 10 52 40 10

30. How many: An Algorithm s1s1 s2s2 s3s3 t1t1 t2t2 t3t3 We want to count non-intersecting sets of paths from s i to t i. 1 8 1 7 40 1 5 25 0 3 13 62 1 5 24 1 6 30 52 40 10 40 62 30 Let a ij be the number of paths from s i to t j.

31. How many: An Algorithm s1s1 s2s2 s3s3 t1t1 t2t2 t3t3 We want to count non-intersecting sets of paths from s i to t i. 1 1 10 1 8 1 6 30 1 4 16 0 2 6 22 52 40 10 40 62 30 10 30 22 Let a ij be the number of paths from s i to t j.

32. How many: An Algorithm s1s1 s2s2 s3s3 t1t1 t2t2 t3t3 We want to count non-intersecting sets of paths from s i to t i. 52 40 10 40 62 30 10 30 22 = 1728. [Gessel, Viennot] Let a ij be the number of paths from s i to t j. Det This is the number domino tilings!!

33. Proof sketch for two paths: a 11 x a 22 counts what we want + extra stuff. a 12 x a 21 also counts the extra stuff. Therefore (a 11 x a 22 ) - (a 12 x a 21 ) counts real tilings. = # (This is the 2 x 2 determinant!)

34. Domino Tilings Where is a tiling? Do any even exist? How many tilings are there? What does a typical tiling look like? When do we stop our algorithms? Why do we care?

35. Why Mathematicians Care Arctic Circle Theorem [Jockush, Propp, Shor] The Aztec Diamond

36. On the chessboard On the hexagonal lat. “Domino tilings” “Lozenge tilings” (little “cubes”) What about tilings on lattices?

37. Why Mathematicians Care

39. Why do we care? Mathematics: Discover patterns Chemistry, Biology: Estimate probabilities Physics: Count and calculate other functions to study a physical system Nanotechnology: Model growth processes

40. Why Physicists Care “Dimer models”: diatomic molecules adhering to the surface of a crystal. The count (“partition function”) determines: specific heat, entropy, free energy, … What does “nature” compute?

41. Domino Tilings Where is a tiling? Do any even exist? How many tilings are there? What does a typical tiling look like? When do we stop our algorithms? Why do we care?

42. What does a typical tiling look like?

43. “Mix” them up!

44. Markov chain for Lozenge Tilings Repeat :  Pick v in the lattice region;  Add / remove the ``cube’’. at v w.p. ½, if possible. v v

45. Markov chain for Lozenge Tilings v 1. The state space is connected. 2.If we do this long enough, each tiling will be equally likely. 3. How long is “long enough” ? v

46. Domino Tilings Where is a tiling? Do any even exist? How many tilings are there? What does a typical tiling look like? When do we stop our algorithms? Why do we care?

47. When do we stop our algorithms? v v Thm: The lozenge Markov chain is “rapidly mixing.” [Luby, R., Sinclair] 2 n, 3 n n 2, 10 √n, … (exponential) n 2, n log n, n 10, … (polynomial)   3. How long is “long enough” ?

48. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets

49. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a 2 x 2 square;

50. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a 2 x 2 square; Rotate, if possible;

51. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a 2 x 2 square; Rotate, if possible; Otherwise do nothing.

52. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx and a color;

53. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx and a color; Recolor, if possible;

54. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx and a color; Recolor, if possible; Otherwise do nothing.

55. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx v and a bit b;

56. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx v and a bit b; If b=1, try to add v

57. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx v and a bit b; If b=1, try to add v; If b=0, try to remove v;

58. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Pick a vtx v and a bit b; If b=1, try to add v; If b=0, try to remove v; O.w. do nothing.

59. What about other models? Potts model Hardcore model Dimer model Domino tilings k-colorings Independent sets Thm: All of these chains are rapidly mixing.

60. HOWEVER... Three-colorings: The local chain is fast for 3-colorings in 2-d [LRS] but slow for 3-colorings in sufficiently high dimension. [Galvin, Kahn, R, Sorkin], [Galvin, R] Independent Sets The local chain is fast for sparse Ind Sets in 2-d [Luby, Vigoda],…, [Weitz] but slow for dense Ind Sets. [R.]

61. Dense Weighted Independent Sets Sparse Fast Phase Transition Slow

62. Why? n 2 /2   (n 2 /2-n/2)  “Even” “Odd” #R/#B

63. Summary Where How What When Why

64. Summary Where How What When Why Algebra Algorithms Combinatorics

65. Summary Where How What When Why Algebra Algorithms Combinatorics Geometry Probability Physics

66. Summary Where How What When Why Algebra Algorithms Combinatorics Geometry Probability Physics Chemistry Nanotechnology Biology

67. THANK YOU ! Math

68. THANK YOU ! Math+ CS

69. THANK YOU ! Math+ CS + Biology, Chemistry,

70. THANK YOU ! + Physics, Nanotechnology,... Math+ CS + Biology, Chemistry,

71. THANK YOU !